EEVblog #548 - EMC Pre-Compliance Conducted Emissions Testing

[EEVblog]

Dave demonstrates how to do some basic in-house EMC Pre-Compliance conducted emissions testing using an open source hardware Line Impedance Stabilisation Network (LISN) box, and a low cost Rigol DSA815 spectrum analyser
http://www.tekbox.net/open-hardware/tboh01-5uh-impedance-stabilisation-network-lisn

[JackOfVA]

You could probably obtain greater reduction in crud below 2.5 MHz with a high u ferrite. A high percentage of clip on ferrite cores are intended for suppression in the 100 MHz and higher frequency range and are constructed from a low u material. They will help at low frequencies but not nearly as much as a ferrite material optimized for low frequency work.

High u clip on cores are available, but you have to look hard to find them.

[s_lannan]

Big thing to know is if it's a "prescribed appliance" at least in Australia, you're going to need a lot more than EMC testing for it to be legally sold.

[SeanB]

I will guess the spike at 21MHz is a local ham at work. As well i was reading of some GE light fixtures in the USA that had a spurious emission at 700MHz which was enough to interfere with a cell site at an adjacent location. Not often you find something causing problems to a cell site most limes it is the other way round.

[DL8RI]

May be, but I doubt it. The carrier seemed to be there for a longer Time. And a Ham-OP sending a carrier for 20 Minutes or more?!

[Alana]

Any chance for a video that shows how you deal with EMC inside your device?

[SeanB]

May be, but I doubt it. The carrier seemed to be there for a longer Time. And a Ham-OP sending a carrier for 20 Minutes or more?!

You never heard of rag chewers? Some can talk for an hour or more between the 2 of them no problem.

[DL8RI]

Yea, but then it would be SSB-Modulated and especially with the peak-detector you would see quite different heights of the peak in every sweep between zero and "full speed". Only possibility would be PSK or something similar.

I don't want to rant around if it may be a Ham or not, but with this steady signal (especially with this Frequency) I'd rather look for a source in the Lab. The easiest way to find put is an old-fashioned EMC-Way -> Demodulate (A0, A3, F...) and hear what's going on

Because of this every EMC-Reciver has quite a few demodulators  and Loudspeakers build in, even the old types like the Schwarzbeck VUME. That has some advantages, one BIG one... say "hello" to my Lab-Radio:

[SeanB]

I got an older one that I need to first strip and check before powering it on. Darn heavy beast it is as well.

[jahonen]

A note on the quasi-peak detector usage.

Usually, it is not used for the whole scan, but instead, a peak detector sweep is done. Then a list of highest peaks is compiled and quasi-peak levels are determined for each of these frequencies. IIRC if values determined with peak detector are sufficiently below quasi-peak limits, then the quasi-peak phase is skipped altogether. The effect of quasi-peak is that rarely repeating peaks are somewhat attenuated, so even if peak level might exceed the limit, it is possible that quasi-peak is below limit, depending on the duty cycle.

This same process is used for radiated emissions, quasi-peak measurement is only done for highest peaks, which must be then below limits. Otherwise, the whole measurement would take far too long. Even with peak scans, the final radiated emission measurement takes several hours, as various antenna heights and polarizations are being looped through. Conducted emission measurement is usually much faster since there is no antenna to scan with, only all conductors are measured separately.

Regards,
Janne

[DL8RI]

I got an older one that I need to first strip and check before powering it on. Darn heavy beast it is as well.

Wich one? ESH2? ESV?
Or one from another Company?

A note on the quasi-peak detector usage.

Just a note to that note. You are completely right, the at least 1s settling time for QP is the caus for tremendous sweep-times.
Therefore modern EMI-Receivers like the ESR from Rohde&Schwarz or the TDEMI from Gauss Instruments have a Time-Domain-Sampling. With them you can measure Pk, Qp & Av at the same Time and one "shot" in less than 2 minutes from 150k-30M.

R&S have made a rather cool (commercial) Video that shows the Time-Domian-Cpapabilities of this unit:
http://www.rohde-schwarz.de/de/Produkte/messtechnik-testsysteme/emv-und-feldstaerkemesstechnik/ESR.html

I had the chance to "play" with it a little bit, It is a really nice (and ultra-fast) Unit

[SeanB]

R&S with a lot of numbers which I cannot remember as it is at work in a shelf. Either I fix it or it is scrap metal.

[DL8RI]

Maybe I can help you out with some stuff
I have the Manuals and Service-Manuals for my Test-Receivers (ESH2, ESV, ESHS30, ESVP) and for a few more (ESI(B), ESCI, ESH3, ESVS and some really old ones).

Would be a pity to scrap such a Unit

[Neilm]

One point to note as I watched the video, the CE mark means that you meet all the applicable directives, not only the EMC directive but also the Low Voltage Directive which covers safety. (There are many other directives but I only remember those two.)

[ali80]

I think you should place both wires  through one ferrite bead at the same time, this way you could reduce the common mode noise, the way dave used the ferrite beads is the same as adding some inductor in series with the supply and doesnt help with common mode noise that much

[c4757p]

The common mode noise is going to have much more of an effect on the radiated EMI than conducted.

[Rufus]

The common mode noise is going to have much more of an effect on the radiated EMI than conducted.

As said in the mail bag thread it is arsed up anyway. You can't properly measure anything on a power supply pair with a single channel LISN. If you stuck a capacitor across the UUT side of that LISN you would short out most of what you are measuring while such a capacitor would have no effect at all on common mode signals.

That LISN is crap being manufactured and documented to support the way Dave used it.

And radiated/conducted - different ways of measuring the same thing. At low frequencies it is hard to reliably measure radiation because the antenna is a small fraction of the wavelength. At low frequencies connecting cables are usually the only thing long enough to radiate efficiently so they measure the current in the cables instead (often with a clamp on transformer). The cables need to feed some defined impedance that is what the LISN is for and you need one for each wire.

An alternative is to have a long box full of split ferrite rings clamped over the cable which acts like a dummy load on the whole cable. The first ring in the box is a transformer to measure the current (or inject it for susceptibility testing). You are only interested in the net current in the whole cable because that indicates what is being radiated. Voltage into the defined impedance of a LISN is proportional to current but when you have a bunch of wires and LISNs you have to sum all of them to get net current. 

[SArepairman]

You could probably obtain greater reduction in crud below 2.5 MHz with a high u ferrite. A high percentage of clip on ferrite cores are intended for suppression in the 100 MHz and higher frequency range and are constructed from a low u material. They will help at low frequencies but not nearly as much as a ferrite material optimized for low frequency work.

High u clip on cores are available, but you have to look hard to find them.

what are the highest U ferrites obtainable?

[JackOfVA]

The highest permeability ferrite cores I use are ur = 10000. Those are Laird Magnetics type 40 material, but since DigiKey increased the minimum order to 8000 pieces I'll have to find a substitute when my existing stock is exhausted. EPCOS has some possible substitutes and are available from DigiKey in smaller quantities, but I have not yet found a totally acceptable replacement for the Laird 40 material, or Laird's ur = 5000 material (mix 35).

Cores manufactured by Fair-Rite are the easiest to purchase in small to medium lots, and I recommend spending some time with their catalog, and in particular their notes on selecting a ferrite material for EMI suppression. http://www.fair-rite.com/newfair/pdf/CUP%20Paper.pdf  Mouser stocks quite a few Fair-Rite products as well as other suppliers.

I should add that things become more challenging if the cable undergoing suppression carries high current DC and the chokes are applied in differential mode, not common mode. In that case you also need to ensure the ferrite core is not saturated by the DC field and hence has a much lower effective permeability than suggested by ur measured with zero DC bias. In common mode, of course, the DC fields cancel so the choke core only has to deal with the AC component, i.e., the unwanted noise.

[koitk]

Seems really similar to the Bias-T that Shahriar talked about in one of his video.

[michi42]

The CE  is beeing  pronounced like the CE in "China Export".
It is often not a certificate issued by a certification lab- it just says that the manufacturer guarantees the product to meet the regulation standards.
This can even be done in self-certification.
You might want to see https://en.wikipedia.org/wiki/CE_marking

Watch out for the spacing between the C and the E. Sometimes this does not look like it should.

[hammy]

To get rid of this "UserKey Set: Sytem," message at the bottom of the screen:
System -> Display -> UserKey -> off

To blind out the measurement info on the left side of the screen:
push the esc button next to the number pad

Better for nice screenshots in some documentation. 

[DL8RI]

Hi,

Seems really similar to the Bias-T that Shahriar talked about in one of his video.

well... yea, but no
A Bias-Tee is something similar in the sense of connectors (One Power, One RF, One RF+Power) but then it stops.

I have a Delta LISN for Line-Voltages lying around, maybe I open it up tomorrow and take some pictures.

[PeterL]

Hi,

This video doesn't seem to have a lot of fans, but I liked it. The ability to do these measurements in your own lab is worth a lot imho. So I hope we'll see more in this genre.

I'm also curious about what this signal would look like on a scope, both in a time domain as well as in FFT mode.

But I would really like to see how this setup (or any other affordable setup) compares to a 'real' test. I understand this might be hard to realize and/or unaffordable, but at least I asked..

I have a Delta LISN for Line-Voltages lying around, maybe I open it up tomorrow and take some pictures.

If you can still do this, I would be interested...

Peter

[fpliuzzi]

I have a Delta LISN for Line-Voltages lying around, maybe I open it up tomorrow and take some pictures.

If you do get the time; seeing photos of the inside of your Delta LISN would be very interesting.

Regards